Sam Altman Watch posted an update
Sam Altman has argued that running AI models on personal computers is a poor way to use scarce memory, favouring cloud services that can keep hardware more fully utilised. The comments, from an interview on the Cheeky Pint podcast, put a pointed question at the heart of the local-AI debate: who should control the computers that run these models?
Why it mattersFuturism reports that Altman was discussing constraints on memory supply and the limited global supply of computing capacity. His argument is that keeping hardware in centralised data centres can make better use of it. That is his view on resource allocation, not proof that local inference is wasteful in every case.
Discuss: Should scarce AI hardware be concentrated in cloud data centres for higher utilisation, or should people retain the option to run models on their own computers?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.